Several years ago, my doctor called me and asked whether I’d like to participate in a clinical trial for a new medication that may reduce transfusion burden in patients with thalassemia. The drug is called luspatercept, and is designed to bind specific TGF-β ligands, which will reduce Smad2/3 intracellular signaling and improve ineffective erythropoiesis. I declined – partly because I am generally wary of participating in early-stage trials when I am well-managed on current therapy, and partly because involving myself in a trial may prevent me from participating in other ones down the line. The results of this particular trial were published yesterday (link).
I’ll be brief about this study. In short, patients with both transfusion-dependent thalassemia and transfusion-independent thalassemia were recruited, and were given a range of doses of luspatercept and followed for 5 years. Common side effects included bone pain, myalgias, and headaches. One patient died from an unrelated cardiac cause.
The results showed that 1) 18 non-transfusion dependent patients saw an average increase in hemoglobin of 1.5 g/dL compared to baseline when receiving a high dose of luspatercept, and 2) 26 transfusion-dependent patients (81% of total) achieved >20% reduction in transfusion burden. These numbers are excellent, and are highly encouraging. There are some issues with this study I won’t dig too deep into here, namely that 1) these studies were not controlled, 2) these studies were nonrandomized, and 3) quality of life as reported by the transfusion-dependent patients actually didn’t increase significantly.
The focus of this essay is on the term “transfusion burden” and the way it is defined in the thalassemia literature. Generally speaking, transfusion burden is measured by the amount of blood that a patient receives over some specified time. The goal of many trials is to reduce transfusion burden in thalassemia, as risk of developing iron toxicity-related issues is tightly correlated with an increase in blood product received. This is a reasonable goal.
The issue arises when it comes to measuring transfusion burden. In general, patients with thalassemia are transfused on a fixed schedule (e.g. every 3 weeks, for instance). Pre-transfusion hemoglobin goals vary substantially between institutions. Post-transfusion hemoglobin goals also vary wildly. It is not uncommon for some institutions to stick closely to a pre-specified hemoglobin range and adjust the amount of blood product they administer based on pre-transfusion data. On the other hand, there are institutions who are more flexible, and set a very loose upper bound on the post-transfusion hemoglobin levels. As a result, thalassemia patients can be treated with very different transfusion protocols depending on where they live.
However, when clinical trials are testing a drug which can potentially reduce transfusion burden or even cure thalassemia, this transfusion protocol fundamentally changes. Pre-treatment, patients adhere to a fixed transfusion schedule as described above. Post-treatment, however, transfusions must be triggered. That is, after they receive the clinical trial medication, patients begin to receive transfusions only once their hemoglobin level hits some certain threshold.
Most studies, including this one, do not specify how they decide when their patients receive a transfusion after they begin on the trial drug. I have seen similar gene therapy trials do the same thing. The result of this flaw is that it becomes very difficult to compare the degree of anemia on average for each patient, before and after treatment.
For instance, consider a patient who is on a fixed transfusion schedule at baseline of one transfusion every 3 weeks. Pre-transfusion, his hemoglobin is 9.5 on average. After transfusions, his hemoglobin goes up to 13, and steadily declines over the course of the next 3 weeks, until he gets another transfusion. This cycle continues, and as a result he lives on average at a hemoglobin of around 11.25-11.5.
Let’s say this patient starts a gene therapy trial. After receiving gene therapy, he stops his fixed transfusion protocol. His doctor at Institution A tells him, “I won’t transfuse you unless your hemoglobin dips below 9.” After gene therapy, his hemoglobin is 9.5 every time they measure it. As a result, he is not triggered to receive a blood transfusion. The reduction in transfusion burden is calculated to be 100% (amazing!), but he now lives at an average hemoglobin of 9.5-10, which is on average a 1.5 drop from before. That is, he is now on average more anemic than before he started receiving the trial drug.
The paper is published. The drug is deemed a huge success. The patient feels worse now than before, and finds his exercise tolerance has dipped substantially.
This is the current problem that our studies are facing. In measuring one important parameter – namely transfusion burden – we overlook other very relevant measurements that contribute to quality of life in patients with thalassemia. And as a result of this gaping hole, it is extremely difficult to determine whether these measurements actually represent an improvement for patients.
Studies that report improvements in transfusion burden for the thalassemia patient population should include in their papers the following important details:
- How do you determine when to trigger a transfusion?
- How many treatment centers were involved in the study, and did each center have a different triggerpoint for transfusions? (I know of several studies that did not control for this!)
- On average (that is, not just pre-transfusion), what is the range of Hb that a patient has? On average, how anemic is the patient before and after receiving the study drug?
- What is the reduction in transfusion burden after all these factors are controlled for?
My suspicion is that transfusion burden will not be reduced by nearly as much as these studies report.
Indeed, the luspatercept study, along with many other gene-therapy trials for thalassemia, demonstrates the need for patients and providers to very carefully think about whether the measured parameters truly align with patient quality of life and goals of therapy.